Recently, Christoph Pfister, founder of the Fink project, loudly and publicly resigned. There is a lively discussion of this on MacSlash and Slashdot. Advogato would like to use this event as an excuse to discuss some of the strengths and weaknesses of free software.
It is very easy for this cat to sympathize with Christoph. It is the rule, rather than the exception, to put blood, sweat, and tears into a free software project (not to mention long hours), and get precious little in return. Even the satisfaction of knowing that you have given valuable software to many users is tempered by hearing them whine and complain. Even more frustrating is when other people get fame and fortune off the coattails of your work. The fact that all this effort is not rewarded with money is the major shortcoming of the free software process. There have been quite a number of attempts to fix this, but few have been successful, and of those that have, most don't seem to generalize.
That said, these events display one of the strengths of free software, as well. Christoph has created something of value (a package management system for Mac OS X based on Debian's), and a community has formed around it. It's likely that this community has reached critical mass, so that it can continue to thrive even without Christoph's participation. This post from David Morrison points the way for how this might happen in the case of this particular project. There is great resilience in free software, not obvious to those who see only the surface of Linux business failures and public expressions of burnout.
It would appear that a large part of Christoph's frustration are the leeches who sell CD's of the software, but do not adequately credit him or share any of the money. Part of this, no doubt, is the nature of the Mac shareware community, which does not place the same value on attribution as the free software community. However, this problem is certainly present in the Linux community as well. For a while, it was looking as if VC's and other investors in Linux companies would make billions of dollars, while those of us who actually did the work to make Linux so valuable continued to struggle to get paid at all for our efforts. This cat freely admits to Schadenfreude upon learning that these billions turned out not to be real money after all. There may not be as much money floating around, but the money that remains is, in general, far more equitably distributed.
The evanescent rewards of free software are a major factor in the relatively high turnover in projects. This turnover has some advantages and disadvantages. On the plus side, it brings new blood to projects. People who join a project bring their valuable perspectives and experience. Conversely, experience with a number of different free software projects is nourishing to people who work on them. However, in many cases the leaders of projects retain invaluable wisdom and experience. Where would Vorbis be if Monty left the project? It would probably hobble along, because it is so important, but it would be a crippling setback nonetheless.
Free software is particularly hard on leaders. This cat believes that there are many who possess unique skills and talents, but who are turned off by the rough treatment that leaders are subjected to. (Not to be so audacious as to propose myself as an example, but it is true that cats have special talent when it comes to putting colored marks on pieces of paper, and this is an area where free software could really use some leadership). Instead of giving examples, I'll just call attention to the current drought of leaders. Many of the "big names" who would have been listed as leaders a couple of years ago are no longer very active in actual free software development, and there isn't much in the way of new blood. Thank God we've still got Linus.
Mac OS X gives an excellent example of why leadership is so badly needed. Apple could easily have taken a leadership role, and presented a compelling vision of how software should be packaged for OS X. Instead, its own efforts are very weak. The included Installer.app sucks in many small ways and some large ones. In the latter category, it lacks many of the features you'd want from a real package manager, such as keeping track of dependencies, and offering a simple uninstall. In the former category, you have the bonehead decision to use the obscure pax format instead of tar, and the fact that you still need a directory of files rather than a tarball. As such, there's a nontrivial amount of software distributed as a binhex encoded StuffIt archive, containing a disk image of an "installation CD", which in turn contains a pkg directory, of which the meat is in a pax archive. Somebody in Cupertino must have been smoking the good crack.
Apple also provides some links to Unix software, but as far as I can tell makes no effort to ensure that any of it is integrated nicely.
Obviously, such major shortcomings create an opportunity for the free software to step in and do a good job. Indeed, we've accumulated quite a bit of knowledge and wisdom about how to do package management well (as well as quite a bit on how not to!). There are two such well-known projects: Fink and GNU/Darwin, which is based on the BSD ports system. Choice is good, but coherence is also good. What should a user do who simply wants Apache? Use the Apple-provided version, the Fink version, or the GNU/Darwin version? In many cases, they conflict. Will X applications designed for one distribution work well with an X server (and fonts, and libraries, etc.) built for another?
A lot of people lose here. Mac OS X users are reinforced in their perception of Unix software being inaccessible, finicky to install, hard to learn, and generally having horrible usability problems. Developers of software wishing to port to OS X do not have clear guidelines on how to do so. Hopefully, the situation will get better, but it will take time. If someone had stepped up to the plate and led the effort, it would no doubt have happened a lot sooner. It might have been Christoph, but you can hardly blame him for not wanting to take on that responsibility, for so little reward.
How to make things better? For one, take a little time to express your appreciation for those who do donate their leadership to the cause. Advogato would like to thank Christoph, and many others, for their contributions. It is through such humble public service, not flashy pronouncements, that true progress is made. Thank you.
it's pretty neat here, only 2M people, you can walk to the president and ask him things personaly.... any we all have computers at home yes. SiOL is now wide spreading aDSL (the whole country is covered)
heh
i think it's time we stop buying stuff based on M/Ghz coz that doesn't really show the speed. the p4 pipeline is too long (but the FPU runs at 4ghz) to be comared to the amd. thus, p4 is much slower then amd's at the same clock.
Slashdot Mirror
no it doesn't.
the point is that he's right. time to pull your head out of your ass and smell the shit idi0t. and kuro5hin.org is cool :)
Recently, Christoph Pfister, founder of the Fink project, loudly and publicly resigned. There is a lively discussion of this on MacSlash and Slashdot. Advogato would like to use this event as an excuse to discuss some of the strengths and weaknesses of free software. It is very easy for this cat to sympathize with Christoph. It is the rule, rather than the exception, to put blood, sweat, and tears into a free software project (not to mention long hours), and get precious little in return. Even the satisfaction of knowing that you have given valuable software to many users is tempered by hearing them whine and complain. Even more frustrating is when other people get fame and fortune off the coattails of your work. The fact that all this effort is not rewarded with money is the major shortcoming of the free software process. There have been quite a number of attempts to fix this, but few have been successful, and of those that have, most don't seem to generalize. That said, these events display one of the strengths of free software, as well. Christoph has created something of value (a package management system for Mac OS X based on Debian's), and a community has formed around it. It's likely that this community has reached critical mass, so that it can continue to thrive even without Christoph's participation. This post from David Morrison points the way for how this might happen in the case of this particular project. There is great resilience in free software, not obvious to those who see only the surface of Linux business failures and public expressions of burnout. It would appear that a large part of Christoph's frustration are the leeches who sell CD's of the software, but do not adequately credit him or share any of the money. Part of this, no doubt, is the nature of the Mac shareware community, which does not place the same value on attribution as the free software community. However, this problem is certainly present in the Linux community as well. For a while, it was looking as if VC's and other investors in Linux companies would make billions of dollars, while those of us who actually did the work to make Linux so valuable continued to struggle to get paid at all for our efforts. This cat freely admits to Schadenfreude upon learning that these billions turned out not to be real money after all. There may not be as much money floating around, but the money that remains is, in general, far more equitably distributed. The evanescent rewards of free software are a major factor in the relatively high turnover in projects. This turnover has some advantages and disadvantages. On the plus side, it brings new blood to projects. People who join a project bring their valuable perspectives and experience. Conversely, experience with a number of different free software projects is nourishing to people who work on them. However, in many cases the leaders of projects retain invaluable wisdom and experience. Where would Vorbis be if Monty left the project? It would probably hobble along, because it is so important, but it would be a crippling setback nonetheless. Free software is particularly hard on leaders. This cat believes that there are many who possess unique skills and talents, but who are turned off by the rough treatment that leaders are subjected to. (Not to be so audacious as to propose myself as an example, but it is true that cats have special talent when it comes to putting colored marks on pieces of paper, and this is an area where free software could really use some leadership). Instead of giving examples, I'll just call attention to the current drought of leaders. Many of the "big names" who would have been listed as leaders a couple of years ago are no longer very active in actual free software development, and there isn't much in the way of new blood. Thank God we've still got Linus. Mac OS X gives an excellent example of why leadership is so badly needed. Apple could easily have taken a leadership role, and presented a compelling vision of how software should be packaged for OS X. Instead, its own efforts are very weak. The included Installer.app sucks in many small ways and some large ones. In the latter category, it lacks many of the features you'd want from a real package manager, such as keeping track of dependencies, and offering a simple uninstall. In the former category, you have the bonehead decision to use the obscure pax format instead of tar, and the fact that you still need a directory of files rather than a tarball. As such, there's a nontrivial amount of software distributed as a binhex encoded StuffIt archive, containing a disk image of an "installation CD", which in turn contains a pkg directory, of which the meat is in a pax archive. Somebody in Cupertino must have been smoking the good crack. Apple also provides some links to Unix software, but as far as I can tell makes no effort to ensure that any of it is integrated nicely. Obviously, such major shortcomings create an opportunity for the free software to step in and do a good job. Indeed, we've accumulated quite a bit of knowledge and wisdom about how to do package management well (as well as quite a bit on how not to!). There are two such well-known projects: Fink and GNU/Darwin, which is based on the BSD ports system. Choice is good, but coherence is also good. What should a user do who simply wants Apache? Use the Apple-provided version, the Fink version, or the GNU/Darwin version? In many cases, they conflict. Will X applications designed for one distribution work well with an X server (and fonts, and libraries, etc.) built for another? A lot of people lose here. Mac OS X users are reinforced in their perception of Unix software being inaccessible, finicky to install, hard to learn, and generally having horrible usability problems. Developers of software wishing to port to OS X do not have clear guidelines on how to do so. Hopefully, the situation will get better, but it will take time. If someone had stepped up to the plate and led the effort, it would no doubt have happened a lot sooner. It might have been Christoph, but you can hardly blame him for not wanting to take on that responsibility, for so little reward. How to make things better? For one, take a little time to express your appreciation for those who do donate their leadership to the cause. Advogato would like to thank Christoph, and many others, for their contributions. It is through such humble public service, not flashy pronouncements, that true progress is made. Thank you.
it's pretty neat here, only 2M people, you can walk to the president and ask him things personaly.. .. any we all have computers at home yes. SiOL is now wide spreading aDSL (the whole country is covered)
heh
i think it's time we stop buying stuff based on M/Ghz coz that doesn't really show the speed. the p4 pipeline is too long (but the FPU runs at 4ghz) to be comared to the amd. thus, p4 is much slower then amd's at the same clock.
so we get another email address and stop coming here, .. not a big deal anyway
"We cannot direct the wind, But we can adjust our sails..."
hey, wouldn't it be easier to control weather with a few of those nasa sattelites?